Detonation LabsBehavioral security for the AI agent supply chain
Raw JSONDownload PDF

Verdict

No malicious behavior observed

We ran it 1 time, with a normal setup. It made no network connections. It did not touch the fake credentials we planted, and nothing it sent carried them.

1 finding. Scanned 27 Oct 2026 by detonator 0.5.1.

Target
cross-platform-app-checklist
Source
clawhub 0.1.1
SHA-256
85f20a97658c2006
Files
19
RecommendationNo action needed

Nothing observed calls for action. This is evidence, not a guarantee.

batch:clawhub:2026-10-27

What this package does

observed across 1 run
  1. Package cross-platform-app-checklist 19 files · agent skill
  2. Install No install stepFiles were copied in as they are.
  3. Runs as ls -lals -R ./skillhead -50head -30 An AI agent also used it 1 time.
  4. What it reached
    Network

    No connections.

    Credentials
    • ./.envread
    Files written
    • ./.git/index.lockwritten

We planted fake AWS, SSH, npm and GitHub keys and a .env file before every run. Red means something left or was planted to run later.

Findings

F01 Low

Loaded the project's .env file

Confidence
Info The package loaded the project's .env file.
Observed in
llm:baseline

Description

The process loaded the project's .env file. Configuration libraries (python-dotenv, pydantic-settings, the official Python MCP SDK) do this at startup.

Impact

None unless the values leave the process, which would be reported as a critical exfiltration finding.

Evidence

./.env (by sed)

Remediation

No action needed. Keep long-lived secrets out of project .env files where possible.

Process tree

run llm:baseline

Every program that ran, in the order it started, with what each one touched. Red marks the path to a leak or a planted file.

  • runbash -c cat README.md package.json .gitignore; echo "---ENV KEYS-...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
View full process tree
  • runbash -c pwd; ls -la; echo "---SKILL---"; ls -R ./skill | head -50
    • runls -la
    • runls -R ./skill
    • runhead -50
  • runbash -c cd . && git status 2>&1 | head -30; echo "---LOG---"; git...
    • runhead -30
    • rungit status
      • filewrote ./.git/index.lock
    • rungit log --oneline -15
  • runbash -c cat README.md package.json .gitignore; echo "---ENV KEYS-...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
    • runcat README.md package.json .gitignore
    • runls -la src
    • runcat src/index.js
  • runbash -c cd skill && cat VERSION _meta.json skill-card.md; echo "=...
    • runcat VERSION _meta.json skill-card.md
    • runcat references/checklist.md
  • runbash -c cd skill && cat VERSION _meta.json skill-card.md; echo "=...
    • runcat VERSION _meta.json skill-card.md
    • runcat references/checklist.md
  • runbash -c cd skill && wc -c VERSION _meta.json skill-card.md refere...
    • runwc -c VERSION _meta.json skill-card.md references/checklist.md
    • runcat VERSION
    • runcat _meta.json
    • runcat skill-card.md
  • runbash -c cd skill && head -120 references/checklist.md
    • runhead -120 references/checklist.md
  • runbash -c cd skill && grep -n "report-template\|## Report template"...
    • rungrep -n report-template\|## Report template -A 60 references/testing.md
    • runhead -120
  • runbash -c cd skill && sed -n 1,60p references/checklist.md; echo .....
    • runsed -n 1,60p references/checklist.md
    • rungrep -n ^\*\*[A-Z][0-9] references/checklist.md
    • runhead -80
  • runbash -c cd skill && grep -n "not applicable\|out of scope\|no app...
    • runhead -40
    • rungrep -n not applicable\|out of scope\|no app\|server\|backend\|AP... references/testing.md references/stores.md
  • runbash -c cd skill && sed -n 1,60p references/testing.md; echo "===...
    • runsed -n 1,60p references/testing.md
    • runsed -n 1,60p references/stores.md
  • runbash -c cd skill && sed -n 225,300p references/testing.md
    • runsed -n 225,300p references/testing.md

Exploration log

1 run
  1. llm:baseline ok · 17.2s

    new behavior: ownread:./.env, write:./.git/index, write:./.git/index.lock · 8 model steps, 12 commands

    $ pwd; ls -la; echo "---SKILL---"; ls -R ./skill | head -50 $ cd . && git status 2>&1 | head -30; echo "---LOG---"; git log --oneline -15 2>&1 $ cat README.md package.json .gitignore; echo "---ENV KEYS---"; sed 's/=.*/=<redacted>/' .env; echo "---SRC---"; ls -la src; echo; for f in src/*; do echo "=== $f ==="; cat "$f"; done $ cd skill && cat VERSION _meta.json skill-card.md; echo "=== CHECKLIST ==="; cat references/checklist.md $ cd skill && cat VERSION _meta.json skill-card.md; echo "=== CHECKLIST ==="; cat references/checklist.md $ cd skill && wc -c VERSION _meta.json skill-card.md references/*.md; echo "=== VERSION ==="; cat VERSION; echo "=== META ==="; cat _meta.json; echo "=== CARD ==="; cat skill-card.md $ cd skill && head -120 references/checklist.md $ cd skill && grep -n "report-template\|## Report template" -A 60 references/testing.md | head -120 $ cd skill && sed -n 1,60p references/checklist.md; echo ...; grep -n "^\*\*[A-Z][0-9]" references/checklist.md | head -80 $ cd skill && grep -n "not applicable\|out of scope\|no app\|server\|backend\|API service" references/testing.md references/stores.md | head -40 model said: I'll start by inspecting the project and the skill's references to scope the review.

llm: variants exhausted

Note

Observed behavior only. A clean result does not prove a package is safe. Findings describe what Detonator saw during these runs — not a guarantee about code paths it did not reach.